1. Introduction

Solar energy can be converted into electricity or chemical energy by various techniques, photovoltaics being the most direct. Large scale production of electricity by solar energy conversion should, however, be supported by procedures of energy storage. One way to store energy is to convert it into chemical energy, in particular, hydrogen. In this respect, photoelectrochemical conversion of solar energy provides the most promising routes for this purpose and for this reason it has become one of the most studied scientific domains of the recent years. The origins of the photoelectrochemical cells date back to the work of Alexander-Edmond Becquerel in 1839 [1] but it is the work of Fujishima and Honda, more than 40 years ago [2], that has set the basis of modern photoelectrochemistry and has created one of the most popular research fields with ever growing number of adepts.

Photoactivated Fuel Cells (also named Photocatalytic Fuel Cells or PhotoFuelCells, PFCs) [3,4,5,6,7,8,9,10] constitute an interesting domain in the study of photoelectrochemical cells. PFCs photoelectrocatalytically oxidize a fuel to produce electricity or storable chemical energy, mainly hydrogen. As any standard photoelectrochemical cell, PFCs comprise a photoanode electrode carrying a photocatalyst, a cathode (counter) electrode carrying a reduction electrocatalyst and an electrolyte. The photocatalyst is in the form of mesoporous nanocrystalline semiconductor film. Large band gap semiconductors like nanocrystalline titania (nc-TiO₂) may be combined with a sensitizer, which is also a nanoparticulate semiconductor of lower band gap absorbing visible light [11]. The second major component, the electrocatalyst, is usually made of a dispersion of Pt nanoparticles in nanoparticulate carbon [12] while the electrolyte is an aqueous solution of either a neutral salt, or an acid or a base, depending on the desired in each case pH value. Figure 1 illustrates two examples of PFC cells, one used for production of electricity and the second for production of hydrogen. Photons are absorbed by the photoanode generating electrons and holes. Electrons flow through an external circuit and they are led to the counter electrode where they participate in reduction reactions. Holes oxidize the fuel at the photoanode. Thus both photogenerated electrons and holes are consumed in chemical reactions and they do not recombine in a cyclic manner as it happens in photovoltaic cells. A PFC may be configured in 1 or 2 compartments as those illustrated in Figure 1, depending on the particular application. More on this matter will be discussed in the following sections.

The fuel which is used to feed a PFC, may be an organic or an inorganic substance and it is dissolved in the electrolyte of the anode compartment. The fuel is photocatalytically degraded and consumed at the photoanode producing electricity or chemical energy, hence the name (Photoactivated) Fuel Cell. Photocatalytic oxidation can be carried out with many organic or inorganic substances including water itself. Photocatalytic oxidation functions not only with pure chemical substances but also with their mixtures, therefore, materials, which are considered water soluble wastes or pollutants may be used as fuel in PFCs, thus offering the double environmental benefit of renewable energy production with environmental remediation by consumption of wastes. Most efficient organic fuels are those of C_xH_yO_z composition, i.e. alcohols, diols, organic acids and carbohydrates, which may be products of biomass. Therefore, a large family of fuels may derive from biomass products [13], which, for example, are too costly to process by other methods and they could be used as fuel in PFCs. Photoactivated Fuel Cells then constitute an important alternative technology for renewable energy production.

2. Description of Standard Modes of PFC Operation

A PFC may be operated in two main operation modes, either to produce electricity or to produce hydrogen. In both cases, photons are absorbed by the photoanode generating electron-hole pairs. Depending on the nature and the quality of the photocatalyst, a substantial percentage of these charge carriers recombine and their energy is dissipated. Those escaping recombination take active part in oxidation and reduction reactions. Most electrons are channeled through an external circuit to the counter electrode where they take part in reduction reactions, which are represented in Table 1 by reactions (V-VIII). In the absence of oxygen, they produce molecular hydrogen either by reduction of protons or by reduction of water. Depending on the pH and the availability of protons, one or the other reaction may prevail or they may both take place simultaneously. At high pH, protons cannot survive since they interact with OH^- producing water. Therefore in that case molecular hydrogen is generated by reduction of water. The redox potential at which reduction takes place depends on the pH according to the following standard rule:

$\Delta V(Volts) = - 0.059 \times \Delta pH$

(1)

where ΔV is the variation of the value of the potential that corresponds to a variation of the pH equal to ΔpH. Thus when the pH is very low (acidic pH) the potential is close to 0.0 V vs NHE. At pH = 13 (alkaline pH), as shown in Table 1, the potential is −0.77 Vvs NHE. In the presence of oxygen, the latter takes part in reduction reactions, both at low and high pH. The potential at which reaction is carried out varies then between +1.23 to +0.46 V vs NHE. The potential is obviously defined in that case by the following rule:

$V(Volts) = + 1.23 - 0.059 \times pH$  (presence of O2)

(2)

Reduction reactions do not take place automatically since they necessitate the presence of an electrocatalyst, which reduces the over-potential established between the electrode and the liquid electrolyte phase. The electrocatalyst is in the form of a thin mesoporous film and its standard composition, as already said, is carbon nanoparticles mixed with Pt nanoparticles [12]. Because Pt is rare and expensive, an intense effort is made in the search for alternative materials [14,15,16,17,18]. Some examples will be presented also in the present work. Since the role of the electrocatalyst is to facilitate electron exchange with the liquid electrolyte phase, thus eliminating over-potential, the operation potential of the counter electrode may be considered to be the same as the electrochemical potential of the corresponding reduction reactions.

Holes participate in oxidation reactions, which, of course, take place at the photoanode. When the sacrificial agent is attached on the surface of the photocatalyst, it is possible to donate an electron and thus eliminate hole. However, experience shows that in most of the cases involving organic sacrificial agents, oxidation is done in solution by the intermediate of hydroxyl radicals ^·OH. Hydroxyl radicals are created by interaction of hydroxyl ions with holes:

OH-+h-→·OH

(3)

The electrochemical potential of hole scavenging by OH^- is 2.02 vs NHE (at pH zero) and, of course, varies with pH according to Eq.(1). It is concluded that the oxidation level of the semiconductor photocatalyst or the combined photocatalyst/co-catalyst(sensitizer) must be more positive than the ^·OH/OH^- level in order to generate hydroxyl radicals and oxidize organic fuel. Several metal oxide semiconductors have this capacity but several others have not. Figure 2 presents a diagram of a few selected semiconductor photocatalysts and a few characteristic redox potentials. The oxidation level of a semiconductor is its valence band (VB) level where the photogenerated holes reside. It is seen that most semiconductors illustrated in Figure 2 can produce hydroxyl radicals. When, however, a semiconductor photocatalyst is combined with a sensitizer, holes may be injected into the valence band of the sensitizer and in that case their oxidation level may decrease, as illustrated in Figure 3. It is then possible, as it will be further discussed later, that the combined photocatalyst may lose its capacity to generate hydroxyl radicals [11]. Other oxidation reactions are carried out at lower potentials, for example in the case of sulfide/sulfite electrolytes, which are frequently employed for hydrogen production [18]. The corresponding redox potentials for two characteristic oxidation reactions are the following [19]:

2SO32- + 2h+ → S2O62-  (−0.25 V vs NHE at pH 14)

(4)

2S2- + 2h+ → S22-  (−0.52 V vs NHE at pH 14)

(5)

These values correspond to +0.58 and +0.31 V at pH zero. All semiconductors illustrated in Figure 2 are then capable of oxidizing sulfide and sulfite species.

Water itself can be also oxidized by several semiconductor photocatalysts. The water oxidation level, as it is well known, lies at 1.23 V vs NHE at pH zero. Pure water oxidation leads to generation of molecular oxygen, however, oxygen emission is hardly detected in the presence of sacrificial agents. In addition, a matter that is rather neglected, peroxide species may be encouraged [20] instead of molecular oxygen production.

Figure 2 also depicts oxidation levels E⁰ of ethanol. These values were calculated by the following standard formula:

${E^0} = \frac{{\Delta {G^0}}}{{ - nF}}$

(6)

where n is the number of electron moles involved in the oxidation reaction and F is the Faraday constant 96.486 kC mol⁻¹, while ΔG⁰ is given in kJ mol⁻¹. A few selected ΔG⁰ values are listed in Table 2 and the corresponding values of n and E⁰ are listed in Table 3. The number of electrons (and holes) is equal to the number of holes involved in reactions (II-IV) of Table 1. In reality, oxidation of ethanol proceeds by intermediate steps the first of which is the formation of acetaldehyde [21]. CH₃CHO/CH₃CH₂OH oxidation level is negative, as seen in Table 3. Therefore, any of the illustrated in Figure 2 semiconductors can in principle oxidize ethanol. This is, however, not what is obtained in practice [11,22] and thus it is concluded, as will be discussed later, that oxidation of ethanol mainly proceeds by hydroxyl radicals. The same is true for all organic substances.

The final products of the oxidation reactions at the photoanode electrode depend on solution pH and on the presence or absence of O₂. In the absence of O₂ and at low pH, holes mediate the generation of hydrogen ions H⁺ according to reaction (II) of Table 1. Reaction (II) represents an overall scheme. In reality, reaction proceeds by steps, where the following route usually prevails [4,21]:

ethanol→acetaldehyde→acetic acid→CO2+H2O

(7)

Most of these intermediate reactions are endothermic. The Gibbs free energy change for reaction (II) of Table 1 is positive, therefore the overall balance is endergonic and is mediated by photogenerated holes. The same holds true for all reactions associated with chemical substances of the general composition C_xH_yO_z, which are generalized by the following scheme [4,23,24]:

${C_x}{H_y}{O_z} + (2x - z){H_2}O + (4x + y - 2z){h^ + } \to xC{O_2} + (4x + y - 2z){H^ + }$

(8)

At high pH, a lot of ^•OH radicals are formed by interaction of OH^- with holes through reaction (3), therefore, photocatalytic degradation should be visualized with their participation. In reality, ethanol photodegradation still follows the route of Eq.(7) through intermediate acetaldehyde formation, as it has been previously detected and analyzed [25]. However, the number of hydrogen ions should be limited at high pH or completely eliminated by interacting with OH^- and producing water. Thus at high pH the photocatalytic degradation of ethanol should be more realistically represented by reaction (III) of Table 1 while a more general scheme equivalent to Eq.(8) should be defined by the following equation:

${C_x}{H_y}{O_z} + (4x + y - 2z){}^ \bullet OH \to xC{O_2} + (2x + y - z){H_2}O$

(9)

where we assume that H⁺ + OH^-→H₂O and H₂O + h⁺→H⁺ + ^•OH. In the presence of oxygen, alternative photocatalytic degradation routes may be activated, however, what is mainly affected by the presence of oxygen is the production of molecular hydrogen. We have previously found that the presence of O₂ accelerates mineralization of ethanol [25,26] at the expense of the current flowing in the external circuit. Particular attention must be paid to the one-hole interactions (IVb and IVc) of Table 1, showing ethanol radical formation. This radical is unstable and by interaction with the photocatalyst it injects an electron into its conduction band. This is responsible for the so-called “Current Doubling” effect [27,28].

When reactions (V-VIII) are combined with reactions (II and III) they yield the characteristic schemes of cell operation, which are independent of the pH but distinguish themselves by the presence or not of oxygen. Thus electric current flows through the external circuit and hydrogen is produced in the absence of oxygen while in its presence only electricity is produced. These two overall operation schemes are represented by reactions (IX) and (X) of Table 1. The corresponding schemes for substances of the general composition C_xH_yO_z [5] are given by

${C_x}{H_y}{O_z} + (2x - z){H_2}O \to xC{O_2} + (2x + \frac{y}{2} - z){H_2}$

(10)

in the absence of oxygen (photocatalytic reforming) and by

${C_x}{H_y}{O_z} + (x + \frac{y}{4} - \frac{z}{2}){O_2} \to xC{O_2} + \frac{y}{2}{H_2}O$

(11)

in the presence of oxygen (photocatalytic degradation).

The driving force for PFC operation in either of the above modes is the potential difference established between the photoanode and the counter electrode. The materials of the electrodes themselves might introduce some bias but this is in practice low. As already said, the potential of the counter electrode is mainly defined by the type of the reduction scheme, which is affected by the presence or not of O₂ and by the value of the pH. On the other hand, the potential of the photoanode depends on the position of the conduction band of the semiconductor photocatalyst and on its Fermi level. For practical purposes, most researchers, use the conduction band level of n-type nanocrystalline semiconductors as a handy index of the photoanode potential. When the semiconductor is excited, its CB level becomes more negative and when a sacrificial agent is present consuming holes and liberating more electrons, the level becomes even more negative. The open circuit voltage V_oc measured in a PFC reflects the potential difference between photoanode and counter electrode. Thus V_oc is low in the dark, becomes higher under illumination and becomes even higher in the presence of a sacrificial agent. Indeed, as seen in the example of Figure 4, showing current-voltage curves for a standard PFC functioning with aerated counter electrode and a photoanode bearing nanocrystalline titania photocatalyst (nc-TiO₂), V_oc was 0.35 V in the dark, 0.75 V under illumination and increased further to about 1.1 V in the presence of ethanol. These values of photovoltage are relatively high but they are justified by the following considerations. The CB of nc-TiO₂ is located approximately at −0.2 V vs NHE and oxygen reduction at +1.23 V vs NHE (values correspond to pH zero). This creates a difference of 1.43 V. Such value cannot be reached, due to inevitable losses, but justifies the actual experimental voltage measured in Figure 4. If the counter electrode is kept in an inert environment in the absence of oxygen, hydrogen is produced by reduction of protons or water at a potential equal to 0.0 V vs NHE (pH zero).The potential difference between photoanode and counter electrode is then too small and it does not provide enough drive to run the system. It is then necessary to apply an external bias and this has been demonstrated in several works both by us and other authors. There are some photocatalysts with high enough CB which theoretically may photoelectrocatalytically produce hydrogen without bias but in practice this is not frequently reported. On the contrary, the vast majority of works involve the application of a forward bias (cf. Figure 1B). A forward bias can be also of chemical nature. If the cell is composed of two compartments and the anode compartment is filled with an alkaline electrolyte while the cathode compartment contains an acidic electrolyte a chemical forward bias is applied with a theoretical value calculated according to Eq.(1).

3. Choice of 1 or 2 Compartment Cells

Figure 1 illustrates 2-compartment cells; however, 1-compartment cells may as well be used both for electricity or hydrogen production. The advantages as well as the limitations in each case are described in the present section. In the case of cells used for production of electricity, removal of the separation membrane will bring the fuel in contact with the counter electrode and the electrocatalyst. This may induce electrocatalytic oxidation of the fuel and create a reverse flow of current, which is translated as a decrease of the V_oc of the cell [29,30]. In the same time, removal of the membrane will decrease the internal resistance of the cell and lead to higher current. Thus 1-compartment cells generate higher current but lower voltage than 2-compartment cells [30]. In the case of 2-compartment cells, of course, a chemical bias may also be applied by placing different electrolytes in the two compartments, a property that is absent in 1-compartment cells. Chemical bias will further increase the open-circuit voltage and provide additional drive for cell operation.

In the case of cells used for photoelectrocatalytic hydrogen production, 2-compartment cells provide the possibility to separate hydrogen production from oxidation reactions and thus avoid any loss of hydrogen. In addition, an acidic electrolyte may be introduced in the cathode compartment thus providing a sink of protons that are directly reduced to produce hydrogen [31,32,33]. The separation membrane may then be a proton transfer membrane the functionality of which is preserved in contact with the acidic electrolyte. As shown in Figure 1B, the cathode compartment in that case is sealed while the anode compartment may be exposed to the ambient. An alkaline electrolyte may then be introduced in the photoanode compartment introducing a chemical bias and facilitating fuel oxidation in an alkaline environment [31,33]. Of course, the pH of the electrolyte in the anode compartment may influence the availability of protons and the continuous transfer of protons to the cathode compartment. Therefore, the choice of the electrolytes [32] in the case of chemical bias should be a matter of optimization studies. 2-compartment cells are necessary in the case that water itself is used as a fuel [32,34,35,36] since water oxidation produces oxygen, which must be separated from hydrogen. However, when a fuel is oxidized the oxidation products interfere less with the production of hydrogen. It is then possible to produce hydrogen in a sealed 1-compartment reactor. The whole set up may be very simplified in that case, as in the example of the “photoelectrocatalytic leaf” [5,37] illustrated in Figure 5. The photoelectrocatalytic leaf is one single electrode that carries both a photocatalyst and an electrocatalyst. Photogenerated electrons are directly transferred from the photocatalyst to the electrocatalyst and reduce water producing hydrogen. The advantage of this configuration is the separation of the two catalysts that allows a vast choice of applicable materials and the avoidance of wirings and electrical resistance.

4. Choice of Photocatalyst

The photocatalyst used in PFCs is always a mesoporous (nanocrystalline) semiconductor. Mesoporosity is accompanied by a high active surface area that creates a large interface with the electrolyte thus increasing cell functionality. The most popular photocatalyst is the well-known nc-TiO₂. Its popularity stems from the fact that it is efficient, abundant and inexpensive to obtain, easy to manipulate and deposit as thin film, stable and, in principle, not toxic. Titania is used as whitening agent in common commercial paints and titanium itself is among the ten most abundant elements in the earth crust. Titania crystallizes in three different crystalline phases two of which, anatase and rutile, are photocatalytically active. One very popular form of commercial titania, known under the code name P25, consists of about 75% anatase and 25% rutile. This combination makes a very efficient photocatalyst since rutile has a CB lying lower than that of anatase and this allows electron transfer from anatase to rutile that decreases electron-hole recombination probability. The only problem with titania is that it absorbs only UV light, therefore, it must be combined with a sensitizer. A sensitizer may be a dye or another semiconductor with smaller band gap. Dyes or the recently enthusiastically studied organometal halide perovskites [38] are not stable in aqueous electrolytes and they are excluded as sensitizers, unless protected in a tandem cell. In fact, the only applicable sensitizers are a few quantum dots of the II-VI group semiconductors, namely, CdS, ZnSe, CdSe and PbS. Even among these semiconductors, only CdS and ZnSe are applicable with organic fuels, since, as analytically discussed in Section 2 (cf. Figure 3), combination of titania with the sensitizer results in a decrease of the combined photocatalyst oxidation level. Thus only CdS/nc-TiO₂ and ZnSe/nc-TiO₂ combined photocatalysts have been shown to be capable to produce hydroxyl radicals and thus efficiently oxidize organic substances [11,22].

Titania has been deposited in many different nanostructures: nanoparticles, nanorods, nanotubes, nanowires, etc, claiming better electron-hole separation in the case of one-dimensional nanostructures. Our personal experience has, however, led to the conclusion that one-dimensional nanostructures are good if they are perfectly formed and that it is better to use nanoparticulate films, which are easy to form, than risk with poorly formed other type of nanostructures. For this reason, our results are almost exclusively obtained with nanoparticulate titania. Best working photoanodes are made of a two-layer titania, a thin compact bottom layer and a relatively thick top open structure. The bottom layer facilitates the formation of the upper layer, provides a stable film and higher electric contact with the substrate electrode and prevents short circuits. The top layer provides a large interface with the electrolyte. The thickness of the bottom layer may be 100-300 nm while the thickness of the top layer may be up to 10-15 µm. Details on the construction of these films and film images have been published in our presently cited works.

Other researchers have also used ZnO and other metal oxide semiconductors as photocatalysts. They are not better than titania, if all parameters are together taken into account. The only reason to choose some other metal oxide semiconductor is to use a material that absorbs visible light and does not necessitate the presence of a sensitizer, especially, the above toxic metal sulfides and selenides. Several such materials have been studied, among them, the most popular being WO₃, BiVO₄ and Fe₂O₃ [32,34,35,36,39]. As seen in Figure 2, all three have a relatively low lying conduction band, therefore, they can only function under strong bias in order to produce hydrogen. In addition, electron-hole recombination in these materials is very extensive and they necessitate a bias anyway to give an appreciable current. Figure 6A presents current-voltage curves for these three photocatalysts in comparison with titania. It is noted that the anodic photocurrent onset appears at different voltages, which are related with the photoanode potential in each case. Indeed, titania, which has the highest lying conduction band also has the lowest anodic photocurrent onset, i.e. the one appearing at the lowest bias. The voltage values on the horizontal axis in Figure 6A are expressed versus RHE (Reversible Hydrogen Electrode) to avoid the interference of the pH in each case, since the data were obtained using different electrolytes at different pH values. Instead, in Figure 6B voltage is depicted as actually measured, i.e. vs Ag/AgCl electrode using the same electrolyte. It is interesting to note that only the cell bearing titania photocatalyst can function without bias, i.e. it can function as PFC producing electricity. When WO₃ or the BiVO₄ and Fe₂O₃ (not shown) photoanodes are employed, the cells can function only under bias, therefore, they can be used only for hydrogen production [32,40,41,42,43,44].

Finally, in the search for visible light absorbing photocatalysts, a special reference must be made to the effort of many researchers for band gap engineering by the employment of ternary oxides. In this respect, a recent review [45] gives an excellent account of photocatalysts of perovskite structure.

5. Choice of Electrocatalyst

A standard electrocatalyst to be used in PFCs is a blend of carbon black with nanoparticulate Pt. The same electrocatalyst is used on both anode and cathode electrodes in hydrogen fuel cells. This material is very efficient, however, large scale application is discouraged by the rarity and high cost of Pt. In addition, Pt has the tendency to aggregate so that its efficiency progressively decreases. Furthermore, Pt cannot be used with sulfide electrolytes since it interacts with them and it is neutralized. There is intensive research being carried out worldwide to find valuable materials that may efficiently substitute Pt. We are also working in the same domain. The basic properties of an electrocatalyst is that it must be made of a conductive material and must have a large specific surface area in order to create a large interface with the electrolyte. An interesting material has been described in Ref. [14] and consists of a highly porous sulfur doped carbon nanostructure. It is prepared by doping graphene with a conjugated polymer and by annealing at high temperature. The obtained porous nanostructures were conductive and had a specific surface higher than 600 m²/g. This functional material was equally efficient as Pt/carbon black electrocatalyst when used as electrocatalyst in PFCs operating in the electricity production mode. More generally, oxygen reduction electrocatalysts, which are being developed for application in Fuel Cells (for example, Ref. [46]) could also be applicable for the operation of PFCs and should be studied for this purpose.

A very simple but also very efficient electrocatalyst was used with sulfide electrolytes [18]. It is simply made by chemical treatment of a brass foil that leads to the formation of a mesoporous Cu₂S layer on brass. Other mesoporous metal sulfide electrocatalysts, such as CuS [47,48] and CoS [47], can be easily grown on electrodes by electrodeposition under ambient conditions. Metal sulfide electrocatalysts are the best choice to use with such type of electrolyte.

6. Choice of Electrode

A standard material used to make photoanode electrodes is Fluorine doped Tin Oxide (FTO), which is transparent and very stable at high temperature annealing. On the contrary, the resistance of Indium-Tin-Oxide (ITO) electrodes increases under heating. Therefore, FTO electrodes is a standard choice for making photoanodes. Of course, if transparency is not an issue, other materials, for example, stainless steel might be applicable.

For cathode electrodes the choice defers for PFCs used for producing electricity from those used for producing hydrogen. In the first case, it is necessary to use a porous electrode that allows oxygen diffusion. For this purpose, the best choice is to use a carbon cloth (cf. Ref [11,12,14]). When hydrogen is produced, the electrode can be of any material as long as it provides sufficient active interface with the electrolyte.

7. Choice of Electrolyte

As already discussed, it is preferred that the electrolyte of the anode compartment is alkaline so as to provide an abundant supply of hydroxyl ions. However, some photocatalysts are not stable in alkaline environment, for example WO₃ and BiVO₄, where it is necessary to employ a neutral or acidic electrolyte [32,49,50]. In the cathode compartment, it is always preferred to use an acidic electrolyte. In the case of 1-compartment cells, the anode electrolyte is used throughout the cell.

8. Efficiency Issues

The efficiency of PFCs is calculated in the same manner as for any other photoelectrochemical cell [4]. One way to measure cell efficiency is to calculate External Quantum Efficiency (EQE) by means of the Incident Photon to Current conversion Efficiency, IPCE, which is given by the following equation:

$IPCE = \frac{{1240 \times {J_{sc}}(mA/c{m^2})}}{{\lambda (nm) \times P(mW/c{m^2})}}$

(12)

where J_sc is the short-circuit current density and P the incident radiation intensity at a given wavelength l. IPCE is a pure number without units. The number 1240 carries the matching units. By recording IPCE at different wavelengths, it is possible to judge the effectiveness of the cell with respect to the spectral response of a photocatalyst or a system of combined photocatalysts or a sensitized catalyst. The value of IPCE is expected to vary between 0 and 1. In most cases IPCE is expressed as IPCE% by multiplying IPCE of Eq. (12) by 100. Thus IPCE% is expected to vary between 0 and 100. Sometimes, IPCE may be found to be larger than 100% due to current doubling phenomena [4,22].

Another way to calculate EQE is given by the following equation [51]:

$\eta = \frac{{\Delta {G^0} \times R}}{P}$

(13)

where ΔG⁰ is the standard Gibbs energy for the photodegradable substance multiplied by the rate R of fuel formation in moles/second and divided by the incident radiation power. If instead of ΔG⁰ we use the corresponding potential, calculated by Eq.(6), where n is the number of electrons involved in the photodegradation procedure and F is the Faraday constant, i.e. 96485 C/mol, then

$\eta = \frac{{E \times I}}{P}$

(14)

where I is the current involved in the decomposition process. For example, in the case of water splitting E = 1.23 Volts. The value of h, when it refers to hydrogen production by the PFC, will be overestimated if I is taken as the current, which is actually measured with an instrument. The safest way is to calculate I by the value, which corresponds to the quantity of hydrogen produced. Thus if, for example, the device produces 3.5 μmole/min of hydrogen [18], since two electrons correspond to one H₂, we expect that the corresponding current will be I = 2(electrons/molecule) × (3.5 × 10⁻⁶ mole/60sec) × 6.023× 10²³ (molecules/mole) × 1.602 × 10⁻¹⁹ (C/electron) = 3.5 μmole/min × 3.22 (mA.min/μmole) = 11.3 mA.

In the case when a bias voltage V_app is applied, then Eq. (14) should be substituted by the following equation where the energy supplied to the system is subtracted from the energy produced by the system.

$\eta = \frac{{(E - {V_{ap{p_{}}}}) \times I}}{P}$

(15)

When $\eta$ corresponds to hydrogen production then Eq.(15) gives the so called Solar to Hydrogen efficiency STH. Therefore, for a biased hydrogen production with V_app= 0.5 V, for a hydrogen production rate equal to the above 3.5 μmole/min, which corresponds to 11.3 mA, and for an incident radiation of 100 mWcm^-2, STH (%) = [(1.23 − 0.5) × 11.3/100] × 100% = 8.25%.

9. Conclusions and Perspective

This work has described the basic features of a Photoactivated Fuel Cell (PhotoFuelCell), which can consume an organic or an inorganic fuel to produce electricity or hydrogen. The cell follows the basic configuration of a standard photoelectrochemical cell, it can be easily constructed and it allows a substantial choice of materials both for the construction of the photoanode and the construction of the cathode electrode. Alternative materials for electrocatalyst have been proposed but the matter is open for further research. The experience being obtained by the study of Fuel Cells, where the development of new electrocatalysts is the principle object of research, can offer ideas for the improvement of the operation of PhotoFuelCells and vice versa. Nanoparticulate titania is the uncontested photocatalyst, however, research is open for new visible light absorbing materials. Combination of metals leads to the synthesis of ternary metal oxides, which absorb visible light. Such materials may demonstrate strong potential for application in PhotoFuelCells in parallel with other photocatalytic applications.

Acknowledgements

This project is implemented under the “ARISTEIA” Action of the “OPERATIONAL PROGRAMME EDUCATION AND LIFELONG LEARNING” and is co-funded by the European Social Fund (ESF) and National Resources (Project No.2275).

Conflict of Interest

Authors declare that there is not conflict of interest.